Está en la página 1de 6

PERSPECTIVES

OPINION

Is mood chemistry?
Eero Castrn
Abstract | The chemical hypothesis of
depression suggests that mood disorders
are caused by a chemical imbalance in the
brain, which can be corrected by
antidepressant drugs. However, recent
evidence indicates that problems in
information processing within neural
networks, rather than changes in chemical
balance, might underlie depression, and that
antidepressant drugs induce plastic changes
in neuronal connectivity, which gradually lead
to improvements in neuronal information
processing and recovery of mood.

The first antidepressants were discovered by


chance almost 50 years ago, when drugs that
had been developed for other disorders were
found to elevate the mood of psychiatric
patients. Soon after this, drugs with antidepressant activity were shown to increase the
extracellular concentrations of two important
monoamine neurotransmitters in the brain
serotonin (5-hydroxytryptamine or 5-HT) and
noradrenaline by inhibiting their catabolism or reuptake to nerve endings. These
findings were the basis for the monoamine
hypothesis of depression, which proposes that
mood disorders are caused by a deficiency in
serotonin or noradrenaline at functionally
important receptor sites in the brain14 (FIG. 1).
It soon became evident that the
monoamine hypothesis in its original form
could not explain all of the effects of antidepressants5. Therefore, the focus of research
was directed towards the receptors and intracellular signal transduction molecules that are
regulated by antidepressant treatment6,7.
Furthermore, because mood disorders often
run in families, genetic studies have been

searching for the genes that might be associated with these familial disorders, and
researchers hope to uncover a molecule that is
malfunctioning in people with depression8.
These research strategies seem to be based on
an extension of the monoamine hypothesis,
the chemical hypothesis of depression (FIG. 2),
which proposes that mood disorders are
caused by structural or functional changes in
particular molecules in the brain, and that
antidepressants function by counteracting
these molecular changes. Over the last few
decades, the view that depression is produced
by a chemical imbalance in the brain has
become widely accepted among scientists,
clinicians and the public.
However, during the past decade, several
observations indicated that there might be
an alternative hypothesis to the chemical
view of depression. This network hypothesis
proposes that mood disorders reflect problems in information processing within particular neural networks in the brain and that
antidepressant drugs and other treatments
that alleviate depression function by gradually improving information processing
within these networks (FIG. 3). This review
discusses the evidence supporting and contradicting the network hypothesis and the
implications of the network view on drug
development and the treatment of mood
disorders.
The chemical hypothesis

We will soon be celebrating the fiftieth


anniversary of the discovery of antidepressants, although the exact date and place of
the discovery is a matter of dispute (for the
history of the discovery of antidepressants, see

NATURE REVIEWS | NEUROSCIENCE

REF. 9). Iproniazid, a drug registered for the


treatment of tuberculosis, was found to
elevate the mood of patients that received it,
and subsequent studies in patients who were
depressed but did not have tuberculosis
showed its effect as an antidepressant9. Simultaneously and independently, imipramine,
an experimental antihistamine with a tricyclic
structure, was found to have antidepressant
effects. These discoveries revolutionized the
recognition and treatment of mood disorders.
In retrospect, it seems unbelievable that
imipramine was introduced to the market
only several years after its antidepressant
effects were discovered, mainly because the
company producing it was unsure that the
number of patients who would benefit from
antidepressant treatment was sufficiently
high9. This now sounds incredible given the
current estimate that major depression is
the single most expensive disorder faced
by Western societies and that, overall, antidepressants are among the best selling drugs.
Soon after this discovery, imipramine and
iproniazid were found to increase the extracellular concentrations of two important
neurotransmitters serotonin and noradrenaline in the brain, by blocking their
re-uptake back to nerve endings or by
inhibiting the main metabolizing enzyme,
monoamine oxidase, respectively. As drugs
that alleviate depression increase extracellular
monoamine concentrations, it was proposed
that depression might be produced by a
serotonin or noradrenaline deficiency at
functionally important receptor sites in the
brain14 (FIG. 1), a proposal that is now known
as the monoamine hypothesis of depression.
Initially, the idea that a complex psychiatric
disorder such as depression could be produced by biochemical changes was met with
widespread scepticism among psychiatrists
and laymen. Nevertheless, during the last few
decades this hypothesis has strongly influenced views about the pathophysiology of
mood disorders, among not only pharmacologists, but also clinicians, other scientists
and the public4.

VOLUME 6 | MARCH 2005 | 2 4 1

2005 Nature Publishing Group

PERSPECTIVES

Presynaptic

Postsynaptic

Figure 1 | Monoamine hypothesis of mood disorders. a | In the normal brain, monoamine


neurotransmitters (yellow) are released and bind to receptors on the postsynaptic neuron. Transmission is
terminated by re-uptake of the transmitter. b | In depression, the decreased concentration of monoamine at
synaptic sites produces a mood disorder. c | Blockade of the re-uptake sites (grey) increases the
concentration of monoamine neurotransmitters available at receptor sites and restores mood.

The monoamine hypothesis focused the


interest of the pharmaceutical industry on
monoamine metabolism for decades.
Imipramine, which inhibits the re-uptake of
both serotonin and noradrenaline (and various other receptors and enzymes), has now
been largely replaced by a host of molecules
that inhibit the uptake of either serotonin or
noradrenaline more selectively, and iproniazid, which inhibits monoamine oxidase,
the main metabolizing enzyme for monoamines, has given way to subtype-selective
monoamine oxidase inhibitors. Although
this focused drug development effort has
clearly been successful from the point of view
of safety, it has been less successful in terms
of efficacy. Modern antidepressants are no
more effective than the first generation of
drugs that were discovered several decades
ago, and electroconvulsive shock treatment
remains the most effective treatment for
depression5,10,11.
It was recognized early on that several
observations conflict with a simple link
between monoamine concentrations in the
brain and depression11. For example, depletion of dietary tryptophan, which significantly decreases the concentration of serotonin in the brain, produces either no effects
or only a mild dysphoria in healthy volunteers
and does not influence the mood of untreated
patients with depression12,13. More importantly, although the effects of antidepressants
on monoamine metabolism can be seen soon
after administration, it typically takes several
weeks of continued treatment for the clinical
antidepressant response to appear11. The discovery that long-term antidepressant treatment produces adaptive changes in
monoamine receptors and in their coupling
to intracellular signal transduction14 caused
the research focus to shift towards the effects
that long-term antidepressant treatments
have on the concentrations of neuropeptides,

242

growth factors and their receptors, and


intracellular signalling molecules4,6,7,15 (FIG. 2).
As a result of this development, the monoamine hypothesis has evolved to what could
be called a chemical or molecular hypothesis
of depression. This hypothesis presumes that
mood disorders are produced by long-term
changes in the production or activity of molecules in the brain and that antidepressants
function by counteracting these molecular
changes. Motivated by this hypothesis,
researchers are using large-scale DNA microarray searches to look for genes that are up- or
downregulated in depression or by antidepressant treatments, in the hope that the
molecules that are encoded by these genes
might be used as targets in the development
of new antidepressant drugs4,16,17.

It was recognized early on


that several observations
conflict with a simple link
between monoamine
concentrations in the brain
and depression.

principal role of the nervous system is not to


handle chemicals but to store and process
information. Arvid Carlsson, one of the main
architects of the concept of chemical neurotransmission in the brain, stated in his Nobel
lecture,However, it must be recognized that
the brain is not a chemical factory but an
extremely complicated survival machine18.
Although chemical neurotransmitters are
crucial for the transfer of information between neurons, information in the brain is not
stored in a chemical form but is thought to be
processed by the complex interactions of neurons in neural networks19,20. These networks
develop through interactions with the environment, and the neuronal structure of, and
neurotransmission in these networks are constantly being refined through activity-dependent synaptic plasticity to optimally process
and store relevant information21 (BOX 1). So,
disorders of the nervous system, including
depression, might represent disturbances in
the activity-dependent information processing of the brain, rather than in the chemical
balance of signalling molecules.
It should be noted that the chemical and
network hypotheses are not mutually exclusive, but are complementary. As the synthesis
and release of several important signalling
molecules are regulated by neuronal activity,
changes in the activity of neural networks
produce changes in the concentration of these
signalling molecules. Therefore, although
the initial effects of antidepressants are obviously chemical and are, in most cases, directed
towards the metabolism of monoamines, the
ensuing adaptive changes in the concentrations of those signalling molecules are tightly
linked to the structure of the neural network,
and might be a consequence of the altered
information processing rather than its cause.
According to this view, antidepressants initiate
a self-repair process, whereby plasticity in
neural networks and chemical neurotransmission indivisibly cooperate and gradually
bring about mood elevation.
Evidence for the network hypothesis

The network hypothesis

But is this view correct? Is mood chemistry?


Observations that have been made during the
last few years indicate that there might be an
alternative to the chemical view of depression
and the action of antidepressants5. This new
hypothesis, the network hypothesis, proposes
that problems in activity-dependent neuronal
communication might underlie depression,
and that antidepressants might work by
improving information processing in the
affected neural networks (FIG. 3). A key aspect
of the network view is the recognition that the

| MARCH 2005 | VOLUME 6

The evidence that supports the network


hypothesis of depression and antidepressant
action is limited and mostly indirect. Part of
the problem is the lack of appropriate
experimental models of depression. In particular, there are no relevant and widely
accepted in vitro models of what might be
going on in the brain during depression.
Furthermore, methods for direct measurement of changes in neural networks in vivo
are only now being developed and have not
yet been applied to neuropharmacological
research19,22.

www.nature.com/reviews/neuro

2005 Nature Publishing Group

PERSPECTIVES

Stress depression
5-HT
NA

Glutamate

BDNF
BDNF

Cortisol

Antidepressants

TRKB
TRKB

NMDA

P
Ca2+

Lithium
AKT

ROS

GSK3

GR
Energy
capacity
P

BAD

TRKB
TRKB

Neuroplasticity and
cellular resilience

BCL-X

BDNF

P
ROS
Ca2+

BCL2

Ras GTP

CREB

RAF

RSK2

MEK

Cytocrome c

Mitochondrion
_
BCL2

Lithium

Lithium VPA

Genetic and
developmental factors

ERK

Failure of
neuroplasticity
signal

Repeated episodes
illness progression

Cerebrovascular
insufficiency

Figure 2 | The chemical hypothesis of depression. The intracellular pathways that are affected by
mood disorders and antidepressants. AKT, protein kinase B; BAD, BLC-associated death promoter;
BCL2, B-cell leukaemia/lymphoma 2; BCL-X, BCL2-like protein 1; BDNF, brain-derived neurotrophic
factor; CREB, cyclic AMP responsive element binding protein; ERK, mitogen activated protein
kinase 1; GR, glucocorticoid receptor; GSK3, glycogen synthase kinase 3; MEK, ERK kinase;
VPA, valproate; NA, noradrenaline; P, phosphate; RAF, RAF proto-oncogene; ROS, reactive oxygen
species; Ras GTP, Ras GTPase-activating protein; RSK2, ribosomal protien S6 kinase polypeptide 3;
TRKB, neurotrophic tyrosine kinase receptor type B; 5-HT, 5-hydroxytryptamine (serotonin).
Adapted, with permission, from REF. 6 (2001) Macmillan Magazines Ltd.

Monoamines, particularly serotonin,


have a significant role during brain development23. Genetic elimination of the 5-HT1A
receptor produces anxiety-type behaviour in
adult mice, but only when the receptor was
absent during early postnatal development
its absence in adulthood produces no
behavioural effects24. Furthermore, mutations in monoamine oxidase A produce
behavioural alterations in both men and
mice25,26, and disrupt the developmental
organization of thalamocortical inputs and
cortical modules in the brains of rodents23,27.
A similar disruption in thalamocortical
organization has also been produced with
the administration of antidepressants and
monoamine oxidase inhibitors during early
postnatal development27,28. Furthermore,
antidepressant treatment during early postnatal life can produce permanent behavioural disturbances in adult animals 29,30.

It is well known that the plasticity of neural


connectivity in the brain is greater and more
extensive during critical periods of postnatal
development than in adults, and that functional structures that are formed during
critical periods remain relatively stable in
adulthood31 (BOX 1). These data highlight the
effects that serotonin and antidepressant
treatments have on the plasticity of neural
networks, and link the effects of antidepressants with the environmental manipulations
that are known to modulate neural network
formation during development31. Moreover,
they indicate that the effects of the drugs
might be more robust in the brain during
early postnatal development and qualitatively different when compared with the
effects seen in the adult brain, which could
have significant implications for the prescription of antidepressants to children and
pregnant mothers.

NATURE REVIEWS | NEUROSCIENCE

Imaging studies in patients with depression have revealed reduced grey matter
volume in the prefrontal cortex3234 and the
hippocampus3539. Morphological changes in
the hippocampus are associated with, and
might be preceded by, functional deficits,
such as memory impairment35. As the neuronal processes and synapses take up most of
the space in the grey matter, reduced volume
might mean reduced neuronal complexity
and connectivity. To at least some degree,
these morphological alterations seem to be
reversible by antidepressant therapies34,40.
The results of a recent study show that
reduced hippocampal volume is particularly
common in patients with depression who
suffered a childhood trauma41, which indicates that severe stress during a critical developmental period might have lasting effects
on the morphology of the brain. These data
support the hypothesis that mood disorders
are associated with compromised information processing in crucial neural networks,
and that the action of antidepressants might
result in morphological and physiological
reorganization of specific neuronal connections in the brain. Furthermore, imaging and
genetic studies are beginning to elucidate
which neural structures are involved in different mental health disorders and which circuits
might be important targets for successful
medications6.
Perhaps the most important evidence for
the network hypothesis is the recent observation that antidepressants increase the
production of new neurons in the rodent
hippocampus42. Importantly, the increased
neurogenesis that is brought about by
chronic antidepressant treatment correlates
with the behavioural effects produced by
antidepressants43. Newly generated neurons
differentiate over time, and are only mature
enough to participate in information
processing several weeks after their birth44.
The fact that this time course correlates with
the delayed onset of the clinical effects of
antidepressants has created a lot of excitement among neuropharmacologists. In the
hippocampi of rodents that have received
antidepressant treatment, the elimination of
neurons through apoptotic cell death
increases simultaneously with increased
neurogenesis, which indicates that antidepressants might increase neuronal turnover
rather than neurogenesis per se 45. This effect
might be functionally analogous to the
overproduction of neurons that occurs
during the development of the peripheral
nervous system (BOX 1), and indicates that
antidepressants might facilitate optimization of neuronal connectivity by increasing

VOLUME 6 | MARCH 2005 | 2 4 3

2005 Nature Publishing Group

PERSPECTIVES
a Healthy

b Depressed

c During treatment

d Recovered

Figure 3 | The network hypothesis of depression. a | In the healthy brain, information is processed
in partially overlapping neural networks. b | In depression, information processing in some networks
does not function properly. c | Antidepressant treatment enhances connectivity in neural networks.
d | Activity-dependent pruning of synapses selects out and stabilizes the active synapses and networks.

the choice of neurons available for selection


through activity-dependent mechanisms. At
a more subtle level, antidepressant drugs
can enhance the sprouting of axons46 and
dendrites47, and support the morphological
maturation of the newborn neurons 47.
These data indicate that, in addition to their
established function in elevating neuronal
turnover in the dentate gyrus, antidepressants might also stimulate the turnover of
axonal branches and synaptic contacts,
thereby providing more material for activity-dependent selection. It should be noted
that increased synaptic turnover might lead
to significant reorganization of neuronal

244

connectivity without any net change in


synaptic number20. Unfortunately, it is difficult
to quantify synaptic turnover in vivo.
One possible mechanism through which
antidepressants might enhance the plasticity
of neuronal connections in the hippocampus
and cerebral cortex is the activation of neurotrophin signalling48,49. Brain-derived neurotrophic factor (BDNF), which is produced
and released by neurons in an activitydependent manner50, has been proposed to
be a crucial factor in the selection and stabilization of active synaptic contacts51 (BOX 1).
Both antidepressants and electroconvulsive
shocks increase the expression and signalling

| MARCH 2005 | VOLUME 6

of BDNF in the hippocampus and cortex5254,


and the injection of BDNF into the brain or
overexpression of its receptor in transgenic
mice produces similar behavioural responses
to those typically observed after treatment
with antidepressant drugs55,56. Consistent
with the importance of BDNF in the effects
of antidepressants, transgenic mice with
reduced BDNF expression or signalling fail
to show these characteristic behavioural
responses after the administration of antidepressants54, which indicates that normal
BDNF signalling might be both necessary and
sufficient for a normal antidepressant effect. It
should be emphasized that the crucial point is
not the increased molecular concentrations
of BDNF as such, but the importance of
this neurotrophin as a mediator of activitydependent neuronal plasticity (BOX 1).
In conclusion, the data that are summarized above provide some evidence that antidepressants, through their acute effects on
monoamine metabolism, activate processes
of plasticity, which are thought to gradually
lead to improved information processing in
the neural networks that are involved in
mood regulation. These processes, which
include neurogenesis and selective neural
elimination, arborization and retraction of
axons and dendrites, and synaptic formation
and pruning, are expected to take time to
develop and mature, which is consistent with
the delayed appearance of the clinical effects
of antidepressants.
However, observations have also been
made that seem to be incompatible with the
network hypothesis. Although depletion
of tryptophan the rate-limiting factor of
serotonin synthesis does not influence the
mood of healthy volunteers and untreated
patients with depression12,13,57, it does produce a rapid relapse of depressive symptoms
in about 50% of remitted patients who are
being, or have recently been treated with
serotonin selective antidepressants12,13.
Furthermore, there is a circadian variation
in mood, and sleep deprivation rapidly
improves the mood of patients with depression, albeit temporarily58. These relatively
rapid effects on mood are difficult to reconcile with the view of their production by a
gradual change in the structure of moodelevating neural networks. It is obvious that
more experimental work will be necessary to
test the new model, but the rapid development of neurophysiological methods and
the imaging of neural networks will help us
to gain further insights into the relationship
between the effects of antidepressants and
neural plasticity, which might become a
fruitful area of further research.

www.nature.com/reviews/neuro

2005 Nature Publishing Group

PERSPECTIVES

Box 1 | Activity-dependent refinement of neural networks


During the development of the peripheral nervous system, neurons are produced in excess
and compete to innervate the target tissues61. Neurotrophic factors, which are secreted by
the target tissues (for example, muscles) in limited quantities, and which are required by
innervating neurons for their survival, select from the competing neurons those that best
innervate the target, and the defeated neurons are eliminated by apoptosis. This competitive
process matches the number of neurons and targets and ensures optimal innervation of the
target cells6163. In the CNS, neurons target other neurons and competitive selection takes
place at several levels: among neurons (as in the peripheral nervous system), among axon
branches and among synaptic contacts64. In the brain, the release of target-derived trophic
factors is activity dependent: the innervating neuron must sufficiently stimulate the target
neuron in order to induce the production and release of the trophic factor50,51. Neurons can
also cooperate to increase the release of trophic factors: simultaneously active neurons that
innervate the same target neuron can induce the release of trophic factors through a much
lower level of activity than is required for a single innervating neuron51. This activitydependent cooperative selection of simultaneously active neurons and the elimination of
inactive and incoherent contacts is considered to be crucial in the development of large,
coherently active neural networks: the neurons that fire together, wire together21. Through
these mechanisms, neuronal structure and neurotransmission are optimized to best store
and process relevant information. During critical periods in early postnatal development,
much larger changes in the connectivity and organization of neural networks take place
than is possible in the adult brain31. Nevertheless, even the adult brain still shows significant
plasticity. The crucial event in activity-dependent plasticity is not the formation of neuronal
contacts (which might occur stochastically and in excess) but the activity-dependent
selection and stabilization of those synapses that mediate useful signals, together with the
selective pruning and elimination of those that produce random noise20. Therefore,
neurogenesis and synaptogenesis cannot simply be considered beneficial, and neuronal
death and synapse elimination harmful; what matters is the optimization of the
signaltonoise ratio in the network. Neurotrophins are important in this process through
their selective release from active connections; indiscriminate increases in the levels of
neurotrophic factors or their signalling (as, for example, would be produced by a
neurotrophic factor receptor agonist) is not expected to be beneficial to the network, as both
active and inactive connections would be similarly supported.

The view of mood disorders as problems of


information processing in the brain has
several important implications. As developmental neurobiologists have been investigating activity-dependent plastic processes
for decades, collaboration between neuropharmacologists, developmental neuroscientists and behavioural geneticists should
be encouraged. Recent studies clearly show
that genetic manipulation of neural circuits
and assessment of the consequences through
in vivo recording techniques and behavioural assays might provide an incredible
potential for begining to define the relationship between circuit properties, behavioural
deficits and potential therapeutics23,24.
During development and in adults, training and rehabilitation produce functional and
anatomical changes in neural networks,
which are reflected in the gradual improvement of the rehearsed action. Analogously,
psychotherapy, cognitive behavioural therapy
and other forms of psychological rehabilitation
could also have therapeutic effects on mood
disorders through use-dependent neuronal

plasticity. Therefore, psychological and


pharmacological therapies, electroconvulsive
shock treatment and placebo effects might all
lead to improved information processing and
mood recovery through mechanisms that initiate similar processes of plasticity (FIG. 4). In
this scenario, a combination of drug treatment and psychological rehabilitation would

Pharmacotherapy

Electroconvulsive therapy

Acitivity-dependent plasticity

Future perspectives

be expected to be more beneficial than either


treatment alone, and there is evidence that
this might be the case59. As the network
hypothesis emphasizes the importance of
environmental information in the process of
activity-dependent selection of neurons and
synapses (BOX 1), it predicts that full recovery
would not even be possible with drug treatment alone, but that external stimuli, such as
social communication, would be required to
provide environmental input for the selection of the appropriate network connections
(FIG. 3). The role of environmental stimulation might also be related to the fact that
major depression is typically cyclical and
often self-limiting, and many patients
improve with time even in the absence of
active treatment.
Finally, the effects of antidepressant drugs
on network plasticity in the brain might
explain why they are effective for many
neuropsychiatric disorders, including anxiety,
obsessive-compulsive disorder, eating disorders, chronic pain and tinnitus. It has been
proposed that some of these disorders, particularly chronic pain, might be produced by
aberrant neuronal connectivity 60.
The hypothesis that mood represents a
functional state of neural networks might
sound incompatible with the efforts of rational drug development. However, the data
reviewed above indicate that the antidepressant drugs that have been used successfully
for several decades might function by initiating such plastic processes, apparently
indirectly, by influencing monoamine
metabolism. It is possible that a similar
process could also be initiated through
other pharmacological mechanisms, which
might become the targets of new antidepressants that could help patients who are
resistant to current drugs and only respond
to electroconvulsive shock treatment.

Psychotherapy

Figure 4 | A combinatorial approach for treating depression based on the network hypothesis.
Depression might reflect disturbed information processing in neural networks (left panel). Antidepressant
drugs, electroconvulsive shock and psychotherapy can all induce activity-dependent plasticity, which
gradually leads to the recovery of connectivity in the affected neural networks (right panel).

NATURE REVIEWS | NEUROSCIENCE

VOLUME 6 | MARCH 2005 | 2 4 5

2005 Nature Publishing Group

PERSPECTIVES
Our view of mood disorders and the
action of antidepressants is beginning to
change from a chemical view towards a network hypothesis, in which problems in information processing and neuronal connectivity
in the brain are the central questions.
However, several important issues have yet to
be addressed. How are neural networks
formed and sustained during development
and in adults? Which neuronal connections
are involved, and are the same networks
affected in all patients? How do antidepressants initiate and support the process of plasticity? Does the delayed appearance of clinical antidepressant effects reflect slow
maturation of neuronal connections, and
could these processes be accelerated? In any
case, it is increasingly evident that the activity-dependent plasticity and connectivity of
neural networks must be considered when
designing future strategies of antidepressant
treatment, whether they are pharmacological, psychological or combinatorial.
Eero Castrn is at the Neuroscience Center,
University of Helsinki, Finland.
e-mail: eero.castren@helsinki.fi
doi:1038/nrn1629
1.

2.
3.

4.

5.
6.

7.

8.

9.
10.

11.
12.

13.

14.

15.

16.

17.

246

Bunney, W. E. Jr & Davis, J. M. Norepinephrine in


depressive reactions. A review. Arch. Gen. Psychiatry 13,
483494 (1965).
Coppen, A. The biochemistry of affective disorders. Br.
J. Psychiatry 113, 12371264 (1967).
Schildkraut, J. J. The catecholamine hypothesis of
affective disorders: a review of supporting evidence. Am.
J. Psychiatry 122, 509522 (1965).
Wong, M. L. & Licinio, J. From monoamines to genomic
targets: a paradigm shift for drug discovery in
depression. Nature Rev. Drug Disc. 3, 136150 (2004).
Nestler, E. J. et al. Neurobiology of depression. Neuron
34, 1325 (2002).
Manji, H. K., Drevets, W. C. & Charney, D. S. The cellular
neurobiology of depression. Nature Med. 7, 541547
(2001).
Coyle, J. T. & Duman, R. S. Finding the intracellular
signaling pathways affected by mood disorder
treatments. Neuron 38, 157160 (2003).
Kennedy, J. L., Farrer, L. A., Andreasen, N. C., Mayeux, R.
& George-Hyslop, P. The genetics of adult-onset
neuropsychiatric disease: complexities and conundra?
Science 302, 822826 (2003).
Healy, D. The Antidepressant Era (Harvard Univ. Press,
Cambridge, Massachusetts, 1997).
Duman, R. S. & Vaidya, V. A. Molecular and cellular
actions of chronic electroconvulsive seizures. J. ECT 14,
181193 (1998).
Nestler, E. J. Antidepressant treatments in the 21st
century. Biol. Psychiatry 44, 526533 (1998).
Delgado, P. L. How antidepressants help depression:
mechanisms of action and clinical response. J. Clin.
Psychiatry 65 Suppl. 4, 2530 (2004).
Booij, L., Van der Does, A. J. & Riedel, W. J. Monoamine
depletion in psychiatric and healthy populations: review.
Mol. Psychiatry 8, 951973 (2003).
Sulser, F., Vetulani, J. & Mobley, P. L. Mode of action of
antidepressant drugs. Biochem. Pharmacol. 27,
257261 (1978).
Duman, R. S., Heninger, G. R. & Nestler, E. J.
A molecular and cellular theory of depression. Arch. Gen.
Psychiatry 54, 597606 (1997).
Knuuttila, J. E., Toronen, P. & Castrn, E. Effects of
antidepressant drug imipramine on gene expression in rat
prefrontal cortex. Neurochem. Res. 29, 12351244
(2004).
Newton, S. S. et al. Gene profile of electroconvulsive
seizures: induction of neurotrophic and angiogenic
factors. J. Neurosci. 23, 1084110851 (2003).

18. Carlsson, A. A half-century of neurotransmitter research:


impact on neurology and psychiatry. Nobel lecture.
Nobelprize.org, <http://www.nobel.se/medicine/
laureates/2000/carlsson-lecture.pdf> (2000).
19. Buzsaki, G. Large-scale recording of neuronal
ensembles. Nature Neurosci. 7, 446451 (2004).
20. Hua, J. Y. & Smith, S. J. Neural activity and the dynamics
of central nervous system development. Nature
Neurosci. 7, 327332 (2004).
21. Katz, L. C. & Shatz, C. J. Synaptic activity and the
construction of cortical circuits. Science 274, 11331138
(1996).
22. Varela, F., Lachaux, J. P., Rodriguez, E. & Martinerie, J.
The brainweb: phase synchronization and large-scale
integration. Nature Rev. Neurosci. 2, 229239 (2001).
23. Gaspar, P., Cases, O. & Maroteaux, L. The
developmental role of serotonin: news from mouse
molecular genetics. Nature Rev. Neurosci. 4, 10021012
(2003).
24. Gross, C. et al. Serotonin1A receptor acts during
development to establish normal anxiety-like behaviour in
the adult. Nature 416, 396400 (2002).
25. Brunner, H. G., Nelen, M., Breakefield, X. O., Ropers, H. H.
& van Oost, B. A. Abnormal behavior associated with a
point mutation in the structural gene for monoamine
oxidase A. Science 262, 578580 (1993).
26. Cases, O. et al. Aggressive behavior and altered amounts
of brain serotonin and norepinephrine in mice lacking
MAOA. Science 268, 17631766 (1995).
27. Cases, O. et al. Lack of barrels in the somatosensory
cortex of monoamine oxidase A-deficient mice: role of a
serotonin excess during the critical period. Neuron 16,
297307 (1996).
28. Xu, Y., Sari, Y. & Zhou, F. C. Selective serotonin reuptake
inhibitor disrupts organization of thalamocortical
somatosensory barrels during development. Dev. Brain
Res. 150, 151161 (2004).
29. Ansorge, M. S., Zhou, M., Lira, A., Hen, R. & Gingrich, J. A.
Early-life blockade of the 5-HT transporter alters
emotional behavior in adult mice. Science 306, 879881
(2004).
30. Feng, P., Ma, Y. & Vogel, G. W. The critical window of
brain development from susceptive to insusceptive.
Effects of clomipramine neonatal treatment on sexual
behavior. Brain Res. Dev. Brain Res. 129, 107110
(2001).
31. Berardi, N., Pizzorusso, T. & Maffei, L. Critical periods
during sensory development. Curr. Opin. Neurobiol. 10,
138145 (2000).
32. Bremner, J. D. et al. Reduced volume of orbitofrontal
cortex in major depression. Biol. Psychiatry 51, 273279
(2002).
33. Botteron, K. N., Raichle, M. E., Drevets, W. C., Heath, A. C.
& Todd, R. D. Volumetric reduction in left subgenual
prefrontal cortex in early onset depression. Biol.
Psychiatry 51, 342344 (2002).
34. Drevets, W. C. Neuroimaging and neuropathological
studies of depression: implications for the
cognitiveemotional features of mood disorders. Curr.
Opin. Neurobiol. 11, 240249 (2001).
35. MacQueen, G. M. et al. Course of illness, hippocampal
function, and hippocampal volume in major depression.
Proc. Natl Acad. Sci. USA 100, 13871392 (2003).
36. Sheline, Y. I., Gado, M. H. & Kraemer, H. C. Untreated
depression and hippocampal volume loss. Am.
J. Psychiatry 160, 15161518 (2003).
37. Sheline, Y. I. Neuroimaging studies of mood disorder
effects on the brain. Biol. Psychiatry 54, 338352 (2003).
38. Frodl, T. et al. Hippocampal changes in patients with a
first episode of major depression. Am. J. Psychiatry 159,
11121118 (2002).
39. Mervaala, E. et al. Quantitative MRI of the hippocampus
and amygdala in severe depression. Psychol. Med. 30,
117125 (2000).
40. Drevets, W. C., Bogers, W. & Raichle, M. E. Functional
anatomical correlates of antidepressant drug treatment
assessed using PET measures of regional glucose
metabolism. Eur. Neuropsychopharmacol. 12, 527544
(2002).
41. Vythilingam, M. et al. Childhood trauma associated with
smaller hippocampal volume in women with major
depression. Am. J. Psychiatry 159, 20722080 (2002).
42. Malberg, J. E., Eisch, A. J., Nestler, E. J. & Duman, R. S.
Chronic antidepressant treatment increases
neurogenesis in adult rat hippocampus. J. Neurosci. 20,
91049110 (2000).
43. Santarelli, L. et al. Requirement of hippocampal
neurogenesis for the behavioral effects of
antidepressants. Science 301, 805809 (2003).
44. van Praag, H. et al. Functional neurogenesis in the adult
hippocampus. Nature 415, 10301034 (2002).

| MARCH 2005 | VOLUME 6

45. Sairanen, M., Lucas, G., Ernfors, P., Castrn, M. &


Castrn, E. BDNF and antidepressant drugs have
different but coordinated effects on neuronal turnover,
proliferation and survival in the adult dentate gyrus.
J. Neurosci. 25, 10891094 (2005).
46. Vaidya, V. A., Siuciak, J. A., Du, F. & Duman, R. S.
Hippocampal mossy fiber sprouting induced by chronic
electroconvulsive seizures. Neuroscience 89, 157166
(1999).
47. Fujioka, T., Fujioka, A. & Duman, R. S. Activation of cAMP
signaling facilitates the morphological maturation of
newborn neurons in adult hippocampus. J. Neurosci. 24,
319328 (2004).
48. Altar, C. A. Neurotrophins and depression. Trends
Pharmacol. Sci. 20, 5961 (1999).
49. Castrn, E. Neurotrophic effects of antidepressant drugs.
Curr. Opin. Pharmacol. 4, 5864 (2004).
50. Thoenen, H. Neurotrophins and neuronal plasticity.
Science 270, 593598 (1995).
51. Poo, M. M. Neurotrophins as synaptic modulators.
Nature Rev. Neurosci. 2, 2432 (2001).
52. Nibuya, M., Morinobu, S. & Duman, R. S. Regulation of
BDNF and trkB mRNA in rat brain by chronic
electroconvulsive seizure and antidepressant drug
treatments. J. Neurosci. 15, 75397547 (1995).
53. Russo-Neustadt, A. A., Beard, R. C., Huang, Y. M. &
Cotman, C. W. Physical activity and antidepressant
treatment potentiate the expression of specific brainderived neurotrophic factor transcripts in the rat
hippocampus. Neuroscience 101, 305312
(2000).
54. Saarelainen, T. et al. Activation of the TrkB neurotrophin
receptor is induced by antidepressant drugs and is
required for antidepressant-induced behavioral effects.
J. Neurosci. 23, 349357 (2003).
55. Shirayama, Y., Chen, A. C., Nakagawa, S., Russell, D. S.
& Duman, R. S. Brain-derived neurotrophic factor
produces antidepressant effects in behavioral
models of depression. J. Neurosci. 22, 32513261
(2002).
56. Siuciak, J. A., Lewis, D. R., Wiegand, S. J. & Lindsay, R. M.
Antidepressant-like effect of brain-derived neurotrophic
factor (BDNF). Pharmacol. Biochem. Behav. 56, 131137
(1997).
57. Van der Does, A. J. W. The effects of tryptophan
depletion on mood and psychiatric symptoms. J. Affect.
Disord. 64, 107119 (2001).
58. Wirz-Justice, A. & Van den Hoofdakker, R. H. Sleep
deprivation in depression: what do we know, where do
we go? Biol. Psychiatry 46, 445453 (1999).
59. Treatment for Adolescents With Depression Study
Team. Fluoxetine, cognitive-behavioral therapy, and
their combination for adolescents with depression:
Treatment for Adolescents With Depression Study
(TADS) randomized controlled trial. JAMA 292, 807820
(2004).
60. Wolpaw, J. R. & Tennissen, A. M. Activity-dependent
spinal cord plasticity in health and disease. Annu. Rev.
Neurosci. 24, 807843 (2001).
61. Levi-Montalcini, R. The nerve growth factor: thirty-five
years later. EMBO J. 6, 11451154 (1987).
62. Barde, Y.-A. Trophic factors and neuronal survival.
Neuron 2, 15251534 (1989).
63. Huang, E. J. & Reichardt, L. F. Neurotrophins: roles in
neuronal development and function. Annu. Rev.
Neurosci. 24, 677736 (2001).
64. Cohen-Cory, S. The developing synapse: construction
and modulation of synaptic structures and circuits.
Science 298, 770776 (2002).

Acknowledgments
I would like to thank H. Rauvala, M. Saarma, M. Castrn and
R. Galuske for their comments to the manuscript, and the Sigrid
Juslius Foundation, Sohlberg Foundation and the Academy of
Finland and for support.

Competing interests statement


The author declares no competing financial interests.

Online links
DATABASES
The following terms in this article are linked online to:
Mouse Genome Informatics: http://www.informatics.jax.org/
BDNF | 5-HT1A
FURTHER INFORMATION
Castrns homepage:
http://www.helsinki.fi/neurosci/castren.htm
Access to this interactive links box is free online.

www.nature.com/reviews/neuro

2005 Nature Publishing Group